Hacker News: lsb

New comment by lsb in "Qwen 3.8 27B available on Cerebras at 1500 tokens/s"

lsb — Sun, 06 Sep 2026 22:49:55 +0000

The context window is limited to 64k or 128k. If you’re using this with a coding agent, there’s going to be a lot of compactions. I found that I had a subagent whose compacted context plus prompts and such was over the window and it errored out in opencode.

New comment by lsb in "Markdown Database Pattern"

lsb — Thu, 27 Aug 2026 22:18:00 +0000

As we have more powerful AI plus agents that can run on your phone, keeping everything as a pile of markdown plus some compute on top makes a lot of sense

New comment by lsb in "Qwen3.8-Flash-Next"

lsb — Wed, 26 Aug 2026 14:15:40 +0000

This is glitch art for text, I love it

New comment by lsb in "Three ways to smuggle SQLite into Nix"

lsb — Fri, 21 Aug 2026 23:15:28 +0000

If this much JSON could fit in a microcontroller’s memory (7.5MB in text), and it’s a performance issue, maybe it’s worth upgrading the JSON parser?

New comment by lsb in "Ornith-1.5: From Self-Scaffolding to Self-Improvement"

lsb — Wed, 19 Aug 2026 21:21:57 +0000

The page has comparisons with Qwen 3.6 27b and I’d love to see comparisons with Qwen 3.8 27b, the newer one is much more capable!

New comment by lsb in "Google is making private AI practical with homomorphic encryption"

lsb — Fri, 14 Aug 2026 17:24:03 +0000

Google is making private AI practical with Gemma4 something that you can run without an Internet connection.

All of the proofs of privacy rely on us getting the math right. All of the privacy from unplugging your internet cable is there by default.

New comment by lsb in "Ancient Library – 1,060 Greek/Latin texts, click any word to parse it"

lsb — Fri, 07 Aug 2026 22:49:49 +0000

Fixed!

New comment by lsb in "Ancient Library – 1,060 Greek/Latin texts, click any word to parse it"

lsb — Fri, 07 Aug 2026 19:27:22 +0000

Neat! Similar to my graduate thesis, NoDictionaries: https://nodictionaries.com/cato/de-agri-cultura/156

New comment by lsb in "Who Is America's Homer?"

lsb — Mon, 20 Jul 2026 04:24:02 +0000

Homer is the Homer of America, unless you believe in fencing off universal properties of mankind into exclusive tribal ownership.

(cf Saul Bellow asking “who is the Tolstoy of the Zulus”, and Ralph Wiley in the Atlantic saying that Tolstoy is the Tolstoy of the Zulus.)

New comment by lsb in "Show HN: Visualizing Tiny LLMs from OpenAI's Parameter Golf"

lsb — Thu, 14 May 2026 19:12:36 +0000

Happy to answer any questions :)

Show HN: Visualizing Tiny LLMs from OpenAI's Parameter Golf

lsb — Thu, 14 May 2026 18:52:29 +0000

The two from parameter golf (one I trained, one was the baseline) are just 16MB each! They produce barely plausible English

Comments URL: https://news.ycombinator.com/item?id=48139581

Points: 3

# Comments: 1

New comment by lsb in "Healthchecks.io now uses self-hosted object storage"

lsb — Fri, 17 Apr 2026 15:19:24 +0000

Self Hosted object storage looks neat!

For this project, where you have 120GB of customer data, and thirty requests a second for ~8k objects (0.25MB/s object reads), you’d seem to be able to 100x the throughput vertically scaling on one machine with a file system and an SSD and never thinking about object storage. Would love to see why the complexity

New comment by lsb in "Leanstral: Open-source agent for trustworthy coding and formal proof engineering"

lsb — Mon, 16 Mar 2026 22:26:24 +0000

The real world success they report reminds me of Simon Willison’s Red Green TDD: https://simonwillison.net/guides/agentic-engineering-pattern...

> Instead of taking a stab in the dark, Leanstral rolled up its sleeves. It successfully built test code to recreate the failing environment and diagnosed the underlying issue with definitional equality. The model correctly identified that because def creates a rigid definition requiring explicit unfolding, it was actively blocking the rw tactic from seeing the underlying structure it needed to match.

New comment by lsb in "Lena by qntm (2021)"

lsb — Fri, 13 Feb 2026 06:58:08 +0000

It’s named after the multi-decade data compression test image https://en.wikipedia.org/wiki/Lenna

Buy the book! https://qntm.org/vhitaos

New comment by lsb in "Ask HN: How are you doing RAG locally?"

lsb — Thu, 15 Jan 2026 09:15:15 +0000

I'm using Sonnet with 1M Context Window at work, just stuffing everything in a window (it works fine for now), and I'm hoping to investigate Recursive Language Models with DSPy when I'm using local models with Ollama

New comment by lsb in "Tell HN: No continental US flights due to attack on Venezuela"

lsb — Sat, 03 Jan 2026 09:57:53 +0000

The New York Times has said that the US president has reported capturing the president of Venezuela https://www.nytimes.com/live/2026/01/03/world/trump-united-s...

Source about aviation: primary (I am at an airport now) and also there are no flights going into or out of JFK right now https://www.jfkairport.com/flight-tracker?view=VIEW_DEPARTUR...

Tell HN: No continental US flights due to attack on Venezuela

lsb — Sat, 03 Jan 2026 09:55:19 +0000

Comments URL: https://news.ycombinator.com/item?id=46474744

Points: 11

# Comments: 3

New comment by lsb in "Lite^3, a JSON-compatible zero-copy serialization format"

lsb — Fri, 19 Dec 2025 05:54:42 +0000

This is super interesting!

Apache Arrow is trying to do something similar, using Flatbuffer to serialize with zero-copy and zero-parse semantics, and an index structure built on top of that.

Would love to see comparisons with Arrow

New comment by lsb in "Why are your models so big? (2023)"

lsb — Sat, 06 Dec 2025 00:31:25 +0000

My threshold for “does not need to be smaller” is “can this run on a Raspberry Pi”. This is a helpful benchmark for maximum likely useful optimization.

A Pi has 4 cores and 16GB of memory these days, so, running Qwen3 4B on a pi is pretty comfortable: https://leebutterman.com/2025/11/01/prompt-optimization-on-a...

New comment by lsb in "Show HN: DSPy on a Pi: Cheap Prompt Optimization with GEPA and Qwen3"

lsb — Tue, 18 Nov 2025 22:08:00 +0000

Happy to answer any questions you have :)